Introduction

Unlike motorway tunnels, sewerage tunnels and other infrastructure tunnels, metro tunnels have generally been excavated as twin tunnels. In addition, metro tunnels have a shallow depth. Due to their shallow depth, metro tunnels have generally been constructed in soft soil. For example, independent of the construction method, all metro tunnels in Istanbul have been excavated as twin tunnels, at a shallow depth and generally in soft soils or weak rocks (Ocak 2008a, 2008b, 2009a, 2009b, Ocak and Bilgin 2010). A typical surface settlement basin shape resulting from twin tunnels is seen in Fig. 1.

Fig. 1
figure 1

Surface settlement basin induced by shallow twin tunnel excavation

Full size image

In metro tunnel excavations, it is important to control surface settlements observed before and after excavation that may cause damage to the surface structures. Otherwise, the metro tunnel cannot perform the expected task and the advantages of the metro tunnel are lost (Ocak 2008a). Independent of the excavation method, the interaction between the twin tunnels should be examined and remedial precautions against any damage to existing structures taken prior to construction. Tunneling cost and project time substantially increase due to damage to structures for surface settlements that result above the tolerable limits. In the Otogar–Kirazli metro project, the damage to the environment that was caused by the EPBM excavation increased costs by 15.8 % and the project schedule by 29.3 % (Ocak 2009b).

The basic parameters affecting the ground deformations may be divided into three main groups (Table 1).

Table 1 Main factors affecting the surface settlement
Full size table

Increasing demands on infrastructures increase attention to shallow soft ground tunneling methods in urbanized areas. Many surface and sub-surface structures make underground construction work very delicate due to the influence of ground deformation, which should be definitely limited/controlled to acceptable levels (Ercelebi et al. 2010). For achieving this purpose, compared with traditional tunnel construction techniques, the EPBM has been widely used in metro tunnel construction throughout the world. Recently, surface settlements in EPBM tunneling have been examined and discussed (Chou and Bobet 2002; Chen et al. 2011; Suwansawat 2006; Migliazza et al. 2009). However, the studies done about the field measurements for parallel tunnels in the literature are still insufficient to understand the interaction of surface settlement between two tunnels. There is no systematic field measurement for surface settlement and EPBM parameters conducted in the same site for parallel tunnels.

In this paper, the interaction of longitudinal surface settlements resulting from twin tunnel construction is studied. Extensive instrumentation records, consisting of systematic field measurement for longitudinal surface settlement at the same site for the parallel tunnels of the Otogar–Kirazli metro project, were used in this research. In addition to settlement data, important data pertaining to shield like face pressure (F p) were used. The settlements were analyzed based on the field measurements and shield parameters.

The geology in the study area is composed of fill, sand, very dense sand, clay and hard clay, starting from the surface. The underground water table starts at about 4–5 m below the surface. The tunnel depth varies between 8.8 and 43.2 m in the study area.

General information

General information about the project

The first construction phase of the Istanbul Metro line began in 1992 and was opened to the public in 2000. This line is being gradually extended and additional lines are being constructed in other locations. One of these metro lines is the twin line between Otogar and Kirazli, of 5.77 km. The metro line consists of a 3.87 km tunnel, a 0.62 km cut and cover station and 1.28 km at the grade crossing. The excavation of this section began in May 2006 and was completed in June 2008 (Fig. 2) (Ocak 2009b).

Fig. 2
figure 2

Main route of Otogar–Kirazlı 1 Metro line and location

Full size image

General information about EPBMs and tunnel geometry

Two EPBMs were used for the excavation of the tunnels (Fig. 3), a Herrenknecht EPBM for the left tube and a Lovat EPBM for the right tube. One excavation was kept approximately 100 m behind the other one. Some of the technical features of the machines are summarized in Table 2 (Ocak 2009b).

Fig. 3
figure 3

EPBMs used in the project: a left line, Herrenknecht; b right line, Lovat

Full size image
Table 2 Technical features of the EPBMs (Ocak 2009b)
Full size table

Since the excavated ground bears water and includes stability problems, the excavation chamber was pressurized up to 300 kPa and conditioned by applying water, foam, bentonite and polymers through the injection ports. The chamber pressure was continuously monitored by pressure sensors inside the chamber and auger. An installation of a segment ring of 1.4 m in length (inner diameter of 5.7 m and outer diameter of 6.3 m) and 30 cm in thickness was implemented with a wing-type vacuum erector. The ring was configured as five segments plus a key segment. After installation of the ring, the excavation restarted and the gap between the segment outer perimeter and excavated tunnel perimeter was grouted up to 300 kPa of pressure through the grout channels in the trailing shield (Ocak 2009b).

Site geology

The study area is located between the Otogar and Kirazlı 1 stations in Istanbul (Fig. 2). The geology of the study area is given in Fig. 4 (Ocak 2009b). The relevant geotechnical properties of the layers are summarized in Table 3. The Gungoren formation of Miocene age was found in the study area. The fill layer consisted of sand, clay, gravel and some pieces of masonry. The sand layer, brown at the upper levels and greenish yellow in color at the lower levels, consists of clay, silt and mica. The very dense sand is greenish yellow and contains mica. The clay layer is grayish green in color and includes gravel and sand. The base layer of the tunnel consists of sand and dark green hard clay containing shell. The underground water table is variable, but is generally located 4–5 m below the surface.

Fig. 4
figure 4

Geology of the study area

Full size image
Table 3 Geotechnical properties of the study area
Full size table

Construction parameters of EPBMs

While all the EPBMs were excavated through the tunnel route, a lot of parameters regarding earth pressure balance (EPB) process were collected (Table 4). The F p, the most important parameter for EPB tunneling, changes from 0.5 bar to 2.7 bar for both EPBMs; the average values were 1.63 bar for Lovat and 1.13 bar for Herrenknecht, respectively. Another important parameter affecting surface settlement is the amount of tail grouting. This parameter changed from 4 to 6.5 m3/lining for both EPBMs; the average was a 4.47 m3/lining for Lovat and a 4.14 m3/lining for Herrenknecht, respectively. The tail void was a 2.813 and a 3.732 m3/lining for Lovat and Herrenknecht, respectively. Therefore, the grouting ratio (the ratio between the volume of grouting material and the volume of tail void) changed from 142 to 213 % for Lovat, and from 94 % to 174 % for Herrenknecht. The content of the grouting material used as a backfill is presented in Table 5.

Table 4 Some important parameters pertain to EPBMs during excavation
Full size table
Table 5 Content of grouting material used as backfill
Full size table

Surface settlement measurements on the surface

Contractors performed surface settlements with intervals of approximately 20–30 m, depending on building situations on the surface. For readings of surface settlements, surface monitoring points (SMPs) were installed at suitable points on the surface. The surface settlement measurement readings started before the excavation at that section and continued up to the allowed limits. The accuracy of SMP measurements was typically 0.025 mm per meter for precise leveling. Generally, the surface settlement measurements are performed at five points on the surface including both tunnel sections in Istanbul Metro studies (Fig. 5).

Fig. 5
figure 5

Typical SMP layout above tunnels. a Plan view b A–B section

Full size image

Investigation of longitudinal surface settlement profile

The main source of settlements induced by shield tunneling can be divided into four main factors (ITA/AITES 2007):

  • settlements ahead and above the face,

  • settlements along the shield,

  • settlements at the shield tail skin,

  • settlements due to liner deformations.

A typical shape of a twin tunnel surface settlement is given in Fig. 1. The evolution of the longitudinal profile of the surface settlement in a TBM/EPBM tunnel was previously investigated by various researchers (Pantet 1991; Mroueh and Shahrour 2008; ITA/AITES 2007). However, these investigations could not encompass the effect of the interactions of the twin tunnels. A typical longitudinal surface settlement profile occurring along a shield is shown in Fig. 6. As seen from Fig. 6, the surface settlement occurring during the excavation process can be divided into three parts (ITA/AITES 2007):

Fig. 6
figure 6

A typical longitudinal surface settlement profile along a shield

Full size image

  • settlement on A point (S A): observed by the face intake,

  • settlement on B point (S B): observed along the shield,

  • settlement on C point (S C): observed at the tail skin (Fig. 6).

The important parameters for SMP and EPBMs used in this research are reported in Tables 6 and 7. The SMPs in Tables 6 and 7 are on the first tunnel (right line, RL) and on the second tunnel (left line, LL), respectively. Because of the similarities in the longitudinal surface settlement, only six of the SMP graphs are given in Figs. 7 and 8. While the graphs in Fig. 7 are based on the data in Table 6, those in Fig. 8 are based on the data in Table 7 (on the second tunnel). Both the first tunnel SMP and second tunnel SMPs are situated exactly along the longitudinal axis. The tunnel cover depth (C), rate of tunnel cover depth to tunnel diameter (C/D) and average F p applied during excavation below the monitoring point (F p) are given in Figs. 7 and 8. All figures given in Figs. 7 and 8 are aligned according to C/D values which change from 1.35 to 4.97 and from 1.12 to 4.97 for the first tunnel and second tunnel, respectively.

Table 6 Study points’ properties on the right line (first tunnel)
Full size table
Table 7 Study points’ properties on the right line (second tunnel)
Full size table
Fig. 7
figure 7

Interaction of surface settlements during excavation of the two EPBMs

Full size image
Fig. 8
figure 8

Interaction of surface settlements during excavation of the two EPBMs

Full size image

During the first and second shield advancement for the first tunnel (Fig. 1), the variations in surface settlements measured on six points are shown in Fig. 7. The surface settlements during the first and second shield advancement have similar trends. The surface started to settle when the distance between the first shield and the monitoring section was approximately less than −20 m (about 3D). At some points (SMP12 and SMP03), the surface showed a slight upheaval when there was a close distance between the first shield and the monitoring section, as seen in Fig. 7a and b. The maximum upheaval was 3.3 mm in SMP12 (Fig. 7a). As expected, the peak of the settlement rate occurred when the first shield reached to monitoring section. This peak value continued to rise until the shield passed the monitoring section. After the shield had passed through the monitoring section by approximately 20–30 m, the surface settlement partly stabilized. It can be concluded that, in clay and sand soil, most of the settlement occurred while the cutter advanced from approximately −20 m before the monitoring section to 20 m past the monitoring section, and that subsequent surface settlement due to soil consolidation after the cutter advanced about 20 m through the monitoring section was relatively small. After the first shield operation for the first tunnel, the second shield arrived in the monitoring section in the second tunnel. When the second shield arrived at the monitoring section, the surface started to settle again. However, this settlement effect was smaller than the first shield. On average, the second shield had a 35 % effect on the first tunnel in this study. After the second shield passed through the monitoring section by about 20–30 m, the surface settlement on the first tunnel stabilized. The distribution of the two shields’ S A, S B and S C values for the first tunnel is presented in Table 8.

Table 8 Distribution of two shields’ S A, S B and S C values for the first tunnel
Full size table

During the first and second shield advancement for the second tunnel, the variations in the surface settlements measured at six points are shown in Fig. 8. The surface started to settle, just as in the first tunnel, when the distance between the first shield and the monitoring section was approximately less than −20 m (about 3D). There was no important upheaval during advancement of the first shield. The settlement rate reached an important value when the first shield came to the monitoring section. This settlement value continued to rise until the shield passed the monitoring section. However, on the second tunnel, the first shield had a little effect, 36.8 %. After the first shield had passed through the monitoring section, by about 20–30 m, the surface settlement stabilized. The main important surface settlement occurred when the second shield came to a monitoring point. Due to the disturbance of soil from the first shield, when the second shield arrived at a monitoring point, a very big surface settlement occurred on the second tunnel. After the second shield passed through the monitoring section, by about 20–30 m, the surface settlement on the second tunnel stabilized. The distribution of the two shields’ S A, S B and S C values for the second tunnel is presented in Table 9.

Table 9 Distribution of two shields’ S A, S B and S C values for the second tunnel
Full size table

The magnitude of surface settlement was found to highly depend on the operational parameters, such as the F p used to operate EPB shields. As can be seen from Figs. 7 and 8, at low C/D ratios at close F p high surface settlements occurred (Figs. 7a and 8a). If the C/D ratios increased, surface settlements decreased (Figs. 7b and 8b). But at low F p at the same C/D ratios, high surface settlements occurred (Fig. 8b). At high C/D ratios at close F p, a little surface settlement occurred (Figs. 7c and 8c).

There are different opinions with regard to S A, S B and S C for only one shield in the literature. These percentage values from literature are given in Table 10. As seen from Table 10, the S A, S B and S C values were found to be 38.0, 35.8 and 26.2 %, respectively (the surface settlement values caused by the second shield were eliminated from the measured results).

Table 10 Distribution of surface settlement along the longitudinal axis with EPBM according to various literature
Full size table

This study also showed that F p was a very important parameter affecting the face stability. The change in the surface settlement versus face pressure/necessary face pressure (NFp) is given in Fig. 9. As seen from Fig. 9, if Fp/NFp ratio is close to one, a little surface settlement occurs. If this ratio rises, the surface settlement tends to rise. NFp is computed in a blow formula:

Fig. 9
figure 9

Correlation between Smax and F p/NF p

Full size image

\( P_{\min } = K_{\text{A3}} \cdot 3\sigma_{\text{v}} + P_{\text{w}} \) (Jancsecz and Steiner 1994).

where P min is the necessary minimum face pressure (NF p) for a stable face, K A3 is a coefficient which can be obtained with Fig. 10, σ v is the vertical soil load and P w is the water table load.

Fig. 10
figure 10

Coefficient of K A3 (Jancsecz and Steiner 1994)

Full size image

Conclusions

Thirty longitudinal sections consisting of a lot of SMP monitoring with EPBM tunneling in sand, hard sand, clay and hard clay, some of them below the water table and others above the water table, were used to determine the interactions of the longitudinal surface settlement profiles in shallow twin tunnels. The following conclusions are obtained:

  1. a.

    Two of the most important parameters for the maximum surface settlements are the F p and backfill parameters.

  2. b.

    Good construction quality is very effective in controlling the face stability and minimizing surface settlement.

  3. c.

    In the twin tunnel excavation with EPBM for a longitudinal profile, the settlement rate reached its peak value when the first shield came to the monitoring section for the first tunnel. This peak value continued until the shield passed the monitoring section. After the shield passed through the monitoring section, by about 20–30 m (about 3D), the surface settlement substantially stabilized.

  4. d.

    The second shield’s settlement effect on the first tunnel is less than that of the first shield. On average, the second shield has a 35 % effect on the first tunnel for a longitudinal profile.

  5. e.

    For the second tunnel settlement, the first shield had a slight effect similar to that of the first tunnel settlement. This had a 36.8 % influence on the second tunnel.

  6. f.

    In this study, S A, S B and S C values were found to be 38.0, 35.8 and 26.2 %, respectively for an EPBM (surface settlement values caused by the second shield were eliminated from the measured results).